CN1633814A - Memory-bandwidth efficient FGS encoder - Google Patents
Memory-bandwidth efficient FGS encoder Download PDFInfo
- Publication number
- CN1633814A CN1633814A CNA038039761A CN03803976A CN1633814A CN 1633814 A CN1633814 A CN 1633814A CN A038039761 A CNA038039761 A CN A038039761A CN 03803976 A CN03803976 A CN 03803976A CN 1633814 A CN1633814 A CN 1633814A
- Authority
- CN
- China
- Prior art keywords
- piece
- dct
- bit plane
- bit
- transform block
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/30—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using hierarchical techniques, e.g. scalability
- H04N19/34—Scalability techniques involving progressive bit-plane based encoding of the enhancement layer, e.g. fine granular scalability [FGS]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
- H04N19/129—Scanning of coding units, e.g. zig-zag scan of transform coefficients or flexible macroblock ordering [FMO]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
- H04N19/17—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object
- H04N19/176—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being an image region, e.g. an object the region being a block, e.g. a macroblock
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
- H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
- H04N19/184—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being bits, e.g. of the compressed video stream
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/42—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals characterised by implementation details or hardware specially adapted for video compression or decompression, e.g. dedicated software implementation
- H04N19/423—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals characterised by implementation details or hardware specially adapted for video compression or decompression, e.g. dedicated software implementation characterised by memory arrangements
- H04N19/426—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals characterised by implementation details or hardware specially adapted for video compression or decompression, e.g. dedicated software implementation characterised by memory arrangements using memory downsizing methods
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/60—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding
- H04N19/61—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using transform coding in combination with predictive coding
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Compression, Expansion, Code Conversion, And Decoders (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
A method and apparatus for fine granular scalability encoding. The following steps are repeated (600), for each individual transform block in an image frame. A respective plurality of residual coefficients are decomposed (602) for the respective transform block. A respective plurality of bit-planes (b, b+1) or discrete quantization steps are processed for the respective transform block (400, 401) before decomposing coefficients for a next one of the transform blocks (410, 411) in the image frame.
Description
Invention field
The present invention relates to the realization of a kind of fine and classified (FGS) encoder.
Background of invention
Video stream transmission on Internet protocol (IP) network allows to carry out various multimedia application.The stream transmission of internet video then provides real-time continuous media delivery of content and demonstration, and the scarcity that service quality on the internet (QoS) is guaranteed compensates simultaneously.Because the bandwidth on the IP network and other performance parameters (for example packet loss rate) can change and have unpredictability, therefore in general, the most of streaming solutions that propose at present all are based on certain layering (or gradable) Video Coding Scheme.
There is the gradable method of several videos to be adopted at present by MPEG-2, MPEG-4 and video compression standard H.263 and so on.In these standards, defined the gradable type of time, space and quality (SNR).The scalable video of all these types all comprises basic layer (BL) and an one or more enhancement layer (EL).Usually, the BL of scalable video stream partly represents is to the described stream needed minimal data amount of decoding.The EL part of video flowing is then represented additional information, has therefore strengthened the demonstration of vision signal when receiver is carried out decoding.
Fine and classified property (FGS) is a kind of recently by the selected brand-new video compression framework that is used for the streaming application of MPEG-4 standard.FGS can support the scene that various bandwidth changes, wherein the common network that is based on IP, the especially internet that characterizes of these scenes.Use the image of this class graduated encoding to decode by progressive mode.That is to say that decoder can just be decoded and display image after receiving very small amount of data.Along with decoder receives more multidata, the quality of decoded picture will improve gradually, up to receiving, decode and having shown complete information.In these leading international standards, gradual image encoding is one of pattern supported in the rest image texture coding instrument of JPEG and MPEG-4 video.
EL has used gradual (embedded) codec to compress SNR and time domain remaining data.So, the FGS residual signal is to compress in the mode of bit plane one by one, and wherein this compression begins from the highest significant position plane and finishes (referring to Fig. 1 and 2) in least significant bit planes.
Fig. 1 is one and has shown and pass through entire frame and the diagram of the conventional order of gradual (bit plane connects a bit plane) from highest significant position plane (MSB) 100 to least significant bit planes (LSB) 102 coding.Although only shown an independent interposition plane 101 here, also can encode to any amount of interposition plane.
Fig. 2 is a diagram that has shown the scanning sequency of FSG enhancement layer residue DCT coefficient.This scanning starts from MSB 100 and carries out towards LSB 102.The representative part that in Fig. 2, has only shown bit plane 100 and 101.Each bit plane piece of 8 * 8 200～204,206,210,211,214 all is to use conventional zigzag pattern to scan, and this scanning starts from the upper left corner of piece and ends at the described lower right corner.Use term " bit plane piece " to represent a part of remaining data with the corresponding single bit plane of single piece inside here.
Mode with tetrad (macro block) scans this bit plane piece, begins from the upper left corner and proceeds in clockwise mode.This scanning starts from first bit plane.Connect arrow and then shown described order: after the lower right corner that scans piece 200, scanning proceeds to the upper left corner of piece 201.Scanning begins to proceed to the upper left corner of piece 202 from the lower right corner of piece 201.Scanning begins to proceed to the upper left corner of piece 203 from the lower right corner of piece 202.Scanning begins to proceed to the next macro block that starts from piece 204 upper left corners from the lower right corner of piece 203.After having finished the scanning of first bit plane, will begin second bit plane of same frame scanned for entire frame.More particularly, to each bit plane b=1,2 ..., m, before first BOB(beginning of block) of next bit plane b+1, all piece k=1,2 of bit plane b ..., scan remaining data among the n.
What Fig. 3 showed is a FGS encoder 300 that is used for the prior art of basic layer and enhancement layer.Fig. 3 has shown an example of the function structure of base layer coder 302 and enhancement layer encoder 304.Although the encoding operation that is based on dct transform that Fig. 3 shows also can use other conversion (for example small echo).
Base layer coder 302 comprises a DCT piece 306, a quantize block 308 and an entropy coder 310 that produces part BL stream from initial video.In addition, base layer coder 302 also comprises the motion estimation block 320 that produces two groups of motion vectors from initial video.Wherein one group of motion vector is corresponding to basic frame, and another is organized then corresponding to the time enhancement frame.In addition, also comprised a multiplexer (not shown) that is used for multiplexing base layer motion vector and BL stream here.
As shown in Figure 3, base layer coder 302 also comprises an inverse quantization block 312, inverse DCT piece 314, motion compensation block 316 and a frame memory 318.
As shown in Figure 3, EL encoder 304 comprises a DCT residual image piece 350 that is used to preserve residual image and MC residual image.Residual image then is to be produced by a subtracter 351 that deducts output from the input of quantize block 308.
BL encoder 304 also comprises a memory 352, has wherein comprised the residual image DCT coefficient of decimal format, and a shielding scanning block 354 that is used to shield and scan all FGS bit planes.Comprised the FGS entropy coding piece 356 that produces the FGS enhanced flow by the coding residual image in addition.
In the routine of FGS encoder 300 realizes (referring to Fig. 3), after having carried out dct transform 306, the DCT residual signal is decomposed at several bit planes (from msb to lsb or certain pre-determined bit plane, for example bp_max).
In piece 354, bit plane is scanned in the mode that a bit plane connects a bit plane then, and in piece 356 it is carried out the distance of swimming and VLC coding.Concerning a complete frame, the continuous sweep bit plane needs subsequently the DCT coefficient of preserving in the memory 352 to be conducted interviews.In addition, because the data in the memory 352 are preserved with decimal system mode rather than binary mode (just a bit plane connects a bit plane), therefore visiting the certain bits plane not only needs to obtain corresponding data, but also will use complicated masking operation to extract the bit plane of expectation.
In conventional encoder 300, need to use a memory 352 to preserve the DCT residual coefficient.In addition, for each bit plane, will carry out repeated access to this memory 352 here.And, in piece 354, need to carry out several masking operations in order to obtain the expectation bit plane that to encode.To preserve those and the relevant state information of previous bit plane compression in addition.This is handled needs a large amount of storage access and computing capability.
Therefore, aspect calculating and storage access (bandwidth just), the efficient that the routine of FGS decoder 300 realizes is all very low.
Summary of the invention
The present invention is a kind of method and apparatus that is used to carry out fine granularity scalability.For each independent transform block in the picture frame, following steps will repeat.Also a plurality of corresponding residual coefficients have been decomposed here in addition for corresponding transform block.In the exploded view picture frame before the coefficient of next transform block, for corresponding transform block is handled corresponding a plurality of bit plane or discrete quantization step.
The accompanying drawing summary
Fig. 1 is one and has shown and cross over a complete frame and the conventional order of gradual (bit plane one by one) from MSB to LSB coding.
Fig. 2 is a conventional sweep diagram in proper order that has shown the residue DCT coefficient of FGS enhancement layer.
Fig. 3 is the block diagram of a conventional FGS encoder.
Fig. 4 is one and has shown in the diagram according to the scanning sequency of the residue DCT coefficient of the FGS enhancement layer in the exemplary encoder of the present invention.
Fig. 5 is a block diagram according to exemplary encoder of the present invention.
Fig. 6 is a flow chart that has shown the exemplary method of the residue DCT coefficient of FGS enhancement layer being handled according to the present invention.
Describe in detail
In according to the preferred method of the present invention, no longer be before next of entire frame scanned than the low order plane, to carry out to the scanning that whole bit plane carried out of a whole frame.The substitute is, before the subsequent block of certain frame inside is handled, to each piece carried out complete scanning (from the highest effectively to least significant bit planes or from highest significant position plane to a pre-determined bit plane).
One exemplary embodiment is a kind of to save bandwidth of memory and to avoid the mode of computational complexity to come the alternative approach that the FGS frame is encoded.
The advantage of this new method is:
-the memory that is used for preserving simultaneously all DCT residual coefficients of picture frame no longer is essential;
-significantly reduced the bandwidth access (become and almost can ignore) of carrying out at different bit plane;
-each coefficient only carried out primary shielding is handled rather than be the repeatedly described processing of each bit plane execution;
-needn't preserve the encoding state information of before encoded (just the highest effectively) bit plane;
-FGS coding no longer needs a frame delay that is used for the FGS coding, basic layer and enhancement layer processing more closely can be combined thus, thereby realize higher efficient aspect computational complexity and storage access.
In order to realize this method, here directly the DCT residual coefficient of a complete DCT piece is handled, rather than the bit plane of a whole frame is handled.Below enumerated the false code that is used for general algorithm.
Algorithm
For each the DCT piece k in the image
Directly decompose the DCT residual coefficient in the corresponding positions plane;
The max of computing block k (| DC-coeff|)=Nmax (k)
For each bit plane b less than Nmax (k)
Each bit plane is handled, just it is carried out the distance of swimming and VLC coding
Preserve each bit plane at a diverse location that starts from known location
(if described is not first piece, and then encoded bitplane b is additional to after the b of encoded bitplane of previous piece)
Calculate the maximum among all Nmax of N=(k)
By adding different bit planes, thereby create bit stream through overcompression with sequence of importance (from msb to lsb)
Fig. 4 shows is the scanning sequency of the FGS enhancement layer residue DCT coefficient that is used to handle.Scanning sequency is from conventional sweep sequential update shown in Figure 2 and (yet in case finish scanning, transmission sequence is identical with the transmission sequence of the signal that conventional encoder 300 shown in Figure 3 is exported) of coming.More particularly, on bit plane b, after the upper left corner from bit plane piece 400 scanned the lower right corner, scanning will proceed to the upper left corner of the bit plane piece 401 on the bit plane b+1.Although in Fig. 4, only shown two bit planes (b and b+1), also can have any amount of bit plane.After the lower right corner that scans bit plane piece 401, if there be the 3rd bit plane, then scanning will proceed to the upper left corner of the 3rd first bit plane piece in the bit plane.Only after via each bit plane the bit plane piece 400,401 of first piece being scanned, described scanning just can proceed to the bit plane piece 410 of the piece that is in the second place among the first bit plane b.More generally, for any one piece k, all bit plane b=1,2 ..., the bit plane piece among the n all is to scan before first bit plane piece of scanning block k+1.
Fig. 6 is a flow chart that shows this algorithm.
Circulation of beginning in step 600.To be each independent transform block in the picture frame (for example DCT piece) k repeated execution of steps 602～614 here.
In step 602, the residue DCT coefficient in all bit planes of direct block of decomposition k.That is to say, will come the different bit plane pieces of block of decomposition k in the mode that a bit plane connects a bit plane here, rather than decompose the coefficient of whole bit plane in the mode that a piece connects a piece.
Circulation of beginning in step 604 is wherein to each coefficient repeated execution of steps 606 of piece k.In step 606, the absolute value of quantity (DC coefficient) is calculated.
In step 608, the NMAX (k) of piece (k) is set for abs (DC coefficient) maximum in all coefficients of piece k.
Circulation of beginning for piece k, is each bit plane b repeated execution of steps 612 and 614 in step 610.
Each bit plane to piece k in step 612 is handled, and just it is carried out the distance of swimming and VLC coding.
Each bit plane piece with piece k in step 614 is kept at the position that has nothing in common with each other that starts from known location.For instance, if current block k is not first piece, then the b of encoded bitplane of piece k partly is added on after the b of encoded bitplane of previous piece k-1 (not shown).Therefore, each bit plane that comes the b position in i DCT piece is kept in the position after the position that follows b bit plane in i-1 the DCT piece closely, wherein b be an integer and i be one greater than 1 integer.After each bit plane b has been repeated step 612～614, then be each piece k repeated execution of steps 602～614.Therefore, from the data arrangement of a plurality of bit planes in such compression bit stream, wherein said stream start from one with a plurality of amplitude peaks in a corresponding bit plane of amplitude peak.
In step 616, the sum of bit plane N is set at the maximum NMAX (k) in all pieces.
In step 618, add different bit planes by order (from MSB to LSB), thereby create compression bit stream according to bit plane importance.Preferably, in compression bit stream inside, the data of each bit plane all are arranged in it on the same position that compression bit stream had that Fig. 3 prior art encoder is produced.So just formed a compression bit stream, wherein comprised corresponding a plurality of bit planes of all DCT pieces in the picture frame, wherein the data in the compression bit stream are arranged according to bit plane.Then, this compression bit stream can be decoded by the decoder of the output of the conventional encoder 300 of any Fig. 3 that can decode.
By above-mentioned algorithm, needn't be in decomposable process and the DCT residual signal is deposited in the memory for after a while access.Do not need in addition to carry out continuous masking operation for different bit planes yet.
What Fig. 5 showed is an exemplary FGS encoder 500 that is used for basic layer and enhancement layer.Fig. 5 has shown a function structure example that is used for base layer coder 502 and enhancement layer encoder 504.Although the encoding operation that is based on dct transform that Fig. 5 shows also can use other conversion (for example small echo) here.
As shown in Figure 5, base layer coder 502 comprises a DCT piece 506, a quantize block 508 and an entropy coder 510 that produces part BL stream from initial video.In addition, base layer coder 502 also comprises the motion estimation block 520 that produces two groups of motion vectors from initial video.Wherein one group of motion vector is corresponding to basic frame, and another is organized then corresponding to the time enhancement frame.And wherein also comprised a multiplexer (not shown) that is used for multiplexing base layer motion vector and BL stream.
As shown in Figure 5, base layer coder 502 also comprises an inverse quantization block 512, inverse DCT piece 514, motion compensation block 516 and a frame memory 518.
EL encoder 504 comprises a DCT residual image piece 550 that is used to preserve residual image and MC residual image.Residual image then is to be produced by a subtracter 551 that deducts output from the input of quantize block 508.EL encoder 504 does not need to provide the memory with residue memory function of memory 352 in the prior art EL encoder 304.In addition, EL encoder 504 does not need necessary shielding and the scanning block 354 that is used to shield and scan all FGS bit planes in the prior art EL encoder 304 yet.The substitute is, the bit plane remaining data of each bit plane piece all is directly to be provided to FGS scanning and entropy coding piece 553 from DCT residual image piece 550, and this FGS scanning and entropy coding piece 553 also are included in so that produce the FGS enhanced flow by the decoding residual image.
In the exemplary realization of FGS encoder 500, after having carried out dct transform 506, before continuing next piece, the DCT residual signal of each independent piece (for example piece in the image upper left corner) all be in several bit plane pieces according to a bit plane connect that the mode of a bit plane decomposes continuously (from msb to lsb or certain pre-determined bit plane, bp_max for example), until till the bit plane piece of each bit plane all carried out scanning.
Then, for a kind of realization of compactness, the mode that connects a bit plane with a bit plane in piece 553 scans each piece separately and they is carried out the distance of swimming and VLC coding.Concerning each piece, the residual image data of all bit planes all are to obtain with the binary form that is used for encoding block 553, therefore there is no need to carry out complicated masking operation, in addition, encoding block 553 once only needs whole bit-plane datas of a piece, rather than from the data of the single bit plane of each piece in the frame.Therefore, here do not need in the prior art in order to achieve this end necessary mass-memory unit 352.
The exemplary method and the system that are used for fine granularity scalability have reduced the needed memory of realization FGS encoder, memory bandwidth and have reduced computational complexity.And the link between basic layer and the enhancement layer encoder becomes more tight, thereby allows more effectively to realize the FGS codec by eliminating unnecessary delay and storage.
Method disclosed herein can also be used in combination with FGS coding tools (selecting to strengthen and frequency weighting).Concerning frequency weighting, wherein be that a fixed matrix is applied to entire frame, can after dct transform, carry out displacement immediately thus.Concerning select strengthening, after complete frame was encoded, the displacement of the bit plane of specific macroblock can be carried out immediately or carry out in later phases before the actual scanning of bit plane and VLC coding.Wherein a kind of method in back provides greater flexibility and interactive the selection to strengthen, but exist need be more complicated memory and the defective of flow management.
In addition, except when outside the preceding FGS structure, this mechanism can also be used in the prediction framework of similar MC-FGS (motion compensation---fine and classified) and P-FGS (gradual fine and classified property).Though wherein different processing is used for P-FGS and MC-FGS, texture coding (FGS scanning just and entropy coding) but is identical.Therefore, aforesaidly constructedly can be applied to MC-FGS and P-FGS equally.
Although what exemplary encoder 500 used is dct transform, this method can be used for other conversion equally, for example block-based wavelet coding or match tracing, and even alternative SNR gradability (using discrete quantized step-length rather than bit plane).
The form that the present invention can adopt the processing of being carried out by computer and be used to implement the equipment of these processing is implemented.The present invention can also adopt the form of the computer program code in the tangible medium to be implemented, and wherein said tangible medium can be floppy disk, read-only memory (ROM), CD-ROM, hard disk, high density (" ZIP for example
TM") removable disk driver or other any computer-readable recording mediums, wherein, when computer loaded computer program code and run time version, computer will become one and be used to implement equipment of the present invention.The present invention also can implement by the form that adopts computer program code, for example be kept at the computer program code that loads and/or carry out or pass through some transmission medium transmission in the storage medium, by computer, wherein for instance, described transmission is via electric wire or cable, carries out via optical fiber or via electromagnetic radiation, in computer loading and computer program code, computer will become one and implement equipment of the present invention.In the time of run time version on a general processor, computer program code segments will be configured processor, so that create specific logical circuit.
Although here be to describe of the present inventionly according to one exemplary embodiment, the present invention is not limited thereto.On the contrary, should explain that so that comprise other modification of the present invention and embodiment, those skilled in the art can implement described other modification of the present invention and embodiment under the situation that does not break away from equivalent scope of the present invention to claims widely.
Claims (20)
1. method that is used for fine granularity scalability may further comprise the steps:
(a) be that each independent transform block in the picture frame repeats (600) and carries out following steps:
(i) decompose corresponding a plurality of residual coefficients in (602) corresponding transform block;
(ii) in the exploded view picture frame before the coefficient of next transform block (410,411), to corresponding a plurality of bit planes in the corresponding transform block (400,401) (b, b+1) or the discrete quantized step-length handle (610,612).
2. the process of claim 1 wherein that described transform block is discrete cosine transform (DCT) piece, and described residual coefficient is the DCT residual coefficient.
3. the method for claim 2, wherein, step comprises that (ii) (b, b+1) each bit plane in is carried out the distance of swimming and variable length code (612) to described a plurality of bit planes.
4. the method for claim 2, wherein, step (a) also comprises:
(iii) (b b+1) preserves (614) in different respectively positions with each bit plane.
5. the method for claim 4, wherein, each bit plane that comes the b position is stored in the position after the position of a b bit plane in i-1 the DCT piece that follows hard on described DCT piece in i the DCT piece of described DCT piece, and wherein b is an integer, and i is the integer greater than 1.
6. the method for claim 2 also comprises:
(b) form (618) compression bit streams, (b, b+1), wherein the data in this compression bit stream are arranged according to bit plane wherein to have comprised a plurality of corresponding bit plane of all DCT pieces in the picture frame.
7. the method for claim 6, wherein:
Step (a) also comprises the amplitude peak (NMAX) in any DCT coefficient of definite (608) corresponding DCT piece;
This method also is included in the amplitude peak (N) in definite before (616) a plurality of amplitude peaks of step (b); And
From the data arrangement (618) of described a plurality of bit planes in a compression bit stream, described bit stream start from described a plurality of amplitude peaks in a corresponding bit plane of amplitude peak (b).
8. the method for claim 6, wherein, step (a) and (b) be under the situation that does not require all DCT residual coefficients of preserving picture frame simultaneously, to carry out.
9. the process of claim 1 wherein that (b b+1) comprises each bit plane from highest significant position plane (b) to least significant bit planes (b+1) to described a plurality of bit planes.
10. the process of claim 1 wherein that described transform block forms (506) by one in the group of being made up of the SNR gradability of discrete cosine transform, block-based wavelet transformation or match tracing and use discrete quantized step-length.
11. an equipment (504) that is used to carry out fine granularity scalability comprising:
The device (550) of a plurality of residual coefficients that is used for the independent transform block of exploded view picture frame;
Scanning and code device (553), be used for before the coefficient of the next transform block of exploded view picture frame (410,411), to corresponding a plurality of bit planes in the corresponding transform block (400,401) (b, b+1) or the discrete quantized step-length handle.
12. the equipment of claim 11, wherein, described scanning comprises with code device (553) and is used for scanning the piece of first sequence and the device (610,612,614) that coded data is deposited in second sequence different with first sequence.
13. the equipment of claim 12, wherein:
Described transform block (400,401,410,411) is discrete cosine transform (DCT) piece, and described residual coefficient is the DCT residual coefficient; And
Each bit plane that comes the b position is stored in the position after the position of a b bit plane in i-1 the DCT piece that follows hard on described DCT piece in i the DCT piece of described DCT piece, wherein b be an integer and i be one greater than 1 integer.
14. the equipment of claim 11, wherein, described equipment (504) is not used in the memory of all DCT residual coefficients of storage map picture frame simultaneously.
15. the equipment of claim 11, wherein, described decomposer (550) directly offers described scanning and code device (553) with the residual coefficient data of certain piece, and these residual coefficient data is not kept in the intermediate storage.
16. the equipment of claim 11, wherein, described decomposer (550) is with piece (400,401) residual coefficient data directly offer described scanning and code device (553), and do not come to extract the data of single bit plane (b) by shielding these residual coefficient data from all pieces of picture frame.
17. a computer-readable medium that has on it through calculation of coding machine program code, wherein when processor was carried out this computer program, processor will be carried out a kind of method that is used to carry out fine granularity scalability, comprising following steps:
(a) for repeating (600), each independent transform block in the picture frame carries out following steps:
(i) corresponding a plurality of residual coefficients of decomposition (602) corresponding transform block;
(ii) in the exploded view picture frame before the coefficient of next transform block (410,411), to corresponding a plurality of bit planes in the corresponding transform block (400,401) (b, b+1) or the discrete quantized step-length handle (610,612).
18. the computer-readable medium of claim 17, wherein, described transform block (400,410) is that discrete cosine transform (DCT) piece and described residual coefficient are the DCT residual coefficients.
19. the computer-readable medium of claim 18, wherein, step (ii) comprises carries out the distance of swimming and variable length code (612) to each bit plane in described a plurality of bit planes.
20. the computer-readable medium of claim 18, wherein
Step (a) also comprises each bit plane all is kept at respectively different positions; And
Each bit plane that comes the b position is stored in the position after the position of a b bit plane in i-1 the DCT piece that follows hard on described DCT piece in i the DCT piece of described DCT piece, wherein b be an integer and i be one greater than 1 integer.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US10/076,374 | 2002-02-15 | ||
| US10/076,374 US20030156637A1 (en) | 2002-02-15 | 2002-02-15 | Memory-bandwidth efficient FGS encoder |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN1633814A true CN1633814A (en) | 2005-06-29 |
Family
ID=27732499
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CNA038039761A Pending CN1633814A (en) | 2002-02-15 | 2003-02-05 | Memory-bandwidth efficient FGS encoder |
Country Status (7)
| Country | Link |
|---|---|
| US (1) | US20030156637A1 (en) |
| EP (1) | EP1479246A1 (en) |
| JP (1) | JP2005518163A (en) |
| KR (1) | KR20040083450A (en) |
| CN (1) | CN1633814A (en) |
| AU (1) | AU2003205962A1 (en) |
| WO (1) | WO2003069917A1 (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101159686B (en) * | 2007-11-01 | 2010-04-14 | 北京航空航天大学 | Zero-delayed FGS bandwidth allocation algorithm based on slide window |
| CN111556318A (en) * | 2020-04-24 | 2020-08-18 | 西安万像电子科技有限公司 | Data transmission method and device |
Families Citing this family (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR100556838B1 (en) * | 2002-09-17 | 2006-03-10 | 엘지전자 주식회사 | Fine granularity scalability encoding and decoding apparatus and method |
| EP1680922A1 (en) * | 2003-10-20 | 2006-07-19 | Samsung Electronics Co., Ltd. | Bit-rate control method and apparatus for normalizing visual quality |
| KR100565308B1 (en) * | 2003-11-24 | 2006-03-30 | 엘지전자 주식회사 | Video code and decode apparatus for snr scalability |
| EP1607858A1 (en) * | 2004-06-16 | 2005-12-21 | Koninklijke Philips Electronics N.V. | Bit-plane extraction operation |
| KR100736104B1 (en) * | 2006-04-06 | 2007-07-06 | 삼성전자주식회사 | Video coding method and apparatus supporting independent parsing |
| US20080013624A1 (en) * | 2006-07-14 | 2008-01-17 | Samsung Electronics Co., Ltd. | Method and apparatus for encoding and decoding video signal of fgs layer by reordering transform coefficients |
| WO2013016871A1 (en) * | 2011-08-03 | 2013-02-07 | Mediatek Inc. | Method and video decoder for decoding scalable video stream using inter-layer racing scheme |
| GB2545503B (en) * | 2015-12-18 | 2020-04-08 | Imagination Tech Ltd | Lossy data compression |
Family Cites Families (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5442458A (en) * | 1991-12-18 | 1995-08-15 | Eastman Kodak Company | Method and associated apparatus for encoding bitplanes for improved coding efficiency |
| GB9703441D0 (en) * | 1997-02-19 | 1997-04-09 | British Tech Group | Progressive block-based coding for image compression |
| US6680976B1 (en) * | 1997-07-28 | 2004-01-20 | The Board Of Trustees Of The University Of Illinois | Robust, reliable compression and packetization scheme for transmitting video |
| US6275531B1 (en) * | 1998-07-23 | 2001-08-14 | Optivision, Inc. | Scalable video coding method and apparatus |
| US6263022B1 (en) * | 1999-07-06 | 2001-07-17 | Philips Electronics North America Corp. | System and method for fine granular scalable video with selective quality enhancement |
| US7020342B1 (en) * | 1999-08-27 | 2006-03-28 | Koninklijke Philips Electronics N.V. | Scalable coding |
| US6614936B1 (en) * | 1999-12-03 | 2003-09-02 | Microsoft Corporation | System and method for robust video coding using progressive fine-granularity scalable (PFGS) coding |
| US6456744B1 (en) * | 1999-12-30 | 2002-09-24 | Quikcat.Com, Inc. | Method and apparatus for video compression using sequential frame cellular automata transforms |
| KR100353851B1 (en) * | 2000-07-07 | 2002-09-28 | 한국전자통신연구원 | Water ring scan apparatus and method, video coding/decoding apparatus and method using that |
| AUPR192700A0 (en) * | 2000-12-06 | 2001-01-04 | Canon Kabushiki Kaisha | Storing coding image data in storage of fixed memory size |
| US6798364B2 (en) * | 2002-02-05 | 2004-09-28 | Intel Corporation | Method and apparatus for variable length coding |
-
2002
- 2002-02-15 US US10/076,374 patent/US20030156637A1/en not_active Abandoned
-
2003
- 2003-02-05 AU AU2003205962A patent/AU2003205962A1/en not_active Abandoned
- 2003-02-05 CN CNA038039761A patent/CN1633814A/en active Pending
- 2003-02-05 WO PCT/IB2003/000401 patent/WO2003069917A1/en active Application Filing
- 2003-02-05 EP EP03702845A patent/EP1479246A1/en not_active Withdrawn
- 2003-02-05 KR KR10-2004-7012370A patent/KR20040083450A/en not_active Application Discontinuation
- 2003-02-05 JP JP2003568899A patent/JP2005518163A/en not_active Withdrawn
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101159686B (en) * | 2007-11-01 | 2010-04-14 | 北京航空航天大学 | Zero-delayed FGS bandwidth allocation algorithm based on slide window |
| CN111556318A (en) * | 2020-04-24 | 2020-08-18 | 西安万像电子科技有限公司 | Data transmission method and device |
Also Published As
| Publication number | Publication date |
|---|---|
| EP1479246A1 (en) | 2004-11-24 |
| AU2003205962A1 (en) | 2003-09-04 |
| WO2003069917A1 (en) | 2003-08-21 |
| US20030156637A1 (en) | 2003-08-21 |
| JP2005518163A (en) | 2005-06-16 |
| KR20040083450A (en) | 2004-10-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN1200568C (en) | Optimum scanning method for change coefficient in coding/decoding image and video | |
| CN1303817C (en) | System and method for encoding and decoding residual signal for fine granular scalable video | |
| US6898324B2 (en) | Color encoding and decoding method | |
| CN1196341C (en) | System and method for encoding and decoding enhancement layer data using base layer quantization data | |
| CN1254975C (en) | Method and apparatus for encoding and decoding video data and video signal | |
| CN102144391B (en) | Skip modes for inter-layer residual video coding and decoding | |
| EP1589764A2 (en) | Method and apparatus for supporting motion scalability | |
| CN1722838A (en) | Use the scalable video coding method and apparatus of basal layer | |
| CN1722831A (en) | To comprising basic layer the bit stream pre decoding and the method and apparatus of decoding | |
| CN1893666A (en) | Video encoding and decoding methods and apparatuses | |
| CN101036388A (en) | Method and apparatus for predecoding hybrid bitstream | |
| WO2004015998A1 (en) | System and method for rate-distortion optimized data partitioning for video coding using backward adaptation | |
| US20080089413A1 (en) | Moving Image Encoding Apparatus And Moving Image Encoding Method | |
| CN1640145A (en) | Improved efficiecy fast framework employing higher quality reference frames | |
| US20050152611A1 (en) | Video/image coding method and system enabling region-of-interest | |
| WO2006112642A1 (en) | Method and apparatus for adaptively selecting context model for entropy coding | |
| WO2005086493A1 (en) | Scalable video coding method supporting variable gop size and scalable video encoder | |
| CN1969564A (en) | Method and system for representing correlated set of images | |
| CN1813479A (en) | Video coding in an overcomplete wavelet domain | |
| CN1665299A (en) | Method for designing architecture of scalable video coder decoder | |
| CN1685731A (en) | Scalable video encoding | |
| CN1633814A (en) | Memory-bandwidth efficient FGS encoder | |
| CN1914926A (en) | Moving picture encoding method and device, and moving picture decoding method and device | |
| CN1656816A (en) | Improved efficiency fgst framework employing higher quality reference frames | |
| JP4593720B2 (en) | Method and apparatus for dynamically displaying residue number coefficient |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |